Messages in this thread | ![/](/images/icornerl.gif) | | Date | Wed, 19 Feb 2003 01:11:54 -0300 | From | Werner Almesberger <> | Subject | Re: [RFC] Is an alternative module interface needed/possible? |
| |
Rusty Russell wrote: > Of course, if you wanted to remove the entry at any other time > (eg. hotplug), this doesn't help you one damn bit (which is kind of > your point).
Yep, try_module_get solves the general synchronization problem for the special but interesting case of modules, but not for the general case.
> This is what network devices do, and what the sockopt registration > code does, too, so this is already in the kernel, too. It's not > great, but it becomes a noop for the module deregistration stuff.
Yes, I think just sleeping isn't so bad at all. First of all, we already have the module use count as a kind of "don't unload now" advice (not sure if you plan to phase out MOD_INC_USE_COUNT ?), and second, it's not exactly without precedent anyway. E.g. umount will have little qualms of letting you sleep "forever". (And, naturally, every once in a while, people hate it for this :-)
Anyway, I'll write more about this tomorrow. For tonight, I have my advanced insanity 101 to finish, topic "ptracing more than one UML/TT at the same time".
- Werner
-- _________________________________________________________________________ / Werner Almesberger, Buenos Aires, Argentina wa@almesberger.net / /_http://www.almesberger.net/____________________________________________/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| ![\](/images/icornerr.gif) |